Psychometric Assessment of PROMIS-29 as a Measure of Recovery After Colorectal Surgery: A Prospective Cohort Study

The PROMIS-29 questionnaire assesses general aspects of physical and mental health that may be relevant to surgical recovery. Although this tool has been endorsed by expert consensus for use in perioperative care, evidence regarding its psychometric performance in this context is limited. This study aimed to assess the content validity, internal consistency, construct validity, and responsiveness of PROMIS-29 as a measure of recovery after colorectal surgery.

This study was conducted according to the COnsensus-based Standards for the selection of health Measurement INstruments (COSMIN) guidelines. A cohort of adult patients undergoing elective colorectal resection at two academic hospitals in Canada completed the PROMIS-29 questionnaire preoperatively and weekly until postoperative week (POW) 4. Physical Health Summary (PHS) and Mental Health Summary (MHS) scores were derived from PROMIS-29 t-scores (higher = better health status). Content validity was assessed by comparing PROMIS-29 items against a previously reported conceptual framework of recovery after abdominal surgery. Internal consistency was assessed using Cronbach α. Construct validity was assessed by testing a priori hypotheses that PROMIS-29 scores would be better in patients: (1) with shorter length of stay (≤4 days); (2) without 30-day complications; (3) undergoing minimally invasive surgery; and (4) without a new stoma. Responsiveness was examined by hypothesizing that PROMIS-29 scores would follow the expected recovery trajectory-showing an initial postoperative decline followed by gradual improvement toward preoperative levels. Consistent with PROMIS guidelines, a minimal important difference of 3 points was used as the criterion for confirming hypotheses.

A total of 282 patients were included (mean±SD age 59±16 years; 48% female; 78% minimally invasive; 19% with stoma; median [IQR] hospital stay 3 [1-6] days). PROMIS-29 items covered 36% of health domains considered important by patients in the conceptual framework, indicating limited content validity. Internal consistency across PROMIS domains was acceptable, with Cronbach α ranging from 0.81 to 0.98. Construct validity of PHS and MHS was limited, with 0 of 4 predefined hypotheses supported on POW1 and only 1 to 2 hypotheses supported in subsequent weeks (mean difference between groups ≥3). Responsiveness was supported until POW2 (mean difference between timepoints ≥3), but not after.

Despite demonstrating adequate internal consistency, PROMIS-29 has limited content validity, construct validity, and responsiveness within the first 4 weeks after colorectal surgery. These findings suggest that PROMIS-29 may not fully capture patients' recovery experiences, underscoring the need for psychometrically sound PROMs tailored to the context of perioperative care.